# Error Messages/40558

## Contents

## Error message text

Internal Frontline Solver error: During Function LpObjSa, the objective coefficient lower/upper dual values are not available. In function OptObjectiveSa, the objective coefficient lower/upper dual values are not available. In general, these are only available for a continuous linear program, and not available when you have integer variables or a nonlinear objective or constraint.

## Cause

This error occurs when you attempt to evaluate OptObjectiveSa (in 4.3) or LpObjSa (in 4.2 and earlier) on an optimization problem that contains integer variables, or is nonlinear.

In Analytica 4.2, the error message says "Internal Solver error", but in fact it really shouldn't say that.

## Remedies

If your problem is just that you have integer variables, but your problem is otherwise linear, then you could consider solving the continuous problem and computing the OptObjectiveSa. Although this doesn't guarantee that the same range would apply to the integer problem, in most cases it would probably be pretty close since it means the continuous solution doesn't change over that coefficient range.

You could then use this to validate the range on the integer problem by resolving at the different coefficient boundaries and validating that the solution obtained is the same solution.

Another approach is to vary one objective coefficient and re-solve the integer or nonlinear problem, and then compare to the first solution. In the non-linear case, use the solution found as the guess. If you can bound a coefficient (e.g., using the OptObjectiveSa for the continuous solution), then you can use a binary-search to narrow it down. Because you need to re-solve optimizations repeatedly for each coefficient bound, this approach is computation intensive.

Enable comment auto-refresher